When Should We Be Voting: Why Electoral Reform Should Involve a Referendum

At the end of 2015 the Trudeau government announced that they would be following through on their promise to make the 2015 election the last in Canada to occur under first past the post. They announced that in doing this they would not be putting proposed changes to a popular referendum, but that the government would be conducting public consultations. Referendums are starting to grow in prominence in Canada. The British Columbia government put the Harmonized Sales tax to a referendum in 2011 and in 2015 residents in the greater Vancouver area found themselves voting on a tax to fund a major transit and infrastructure expansion. Three provinces, British Columbia (twice), Ontario, and Prince Edward Island have had referendums on changes to their electoral system. Referendums can be a value tool through which governments can get popular input on policy proposals. They do carry the danger that politicians will abrogate their duty to carefully consider and policy decisions and take responsibility for them. Referendums should be used sparingly, but are important for political decisions that affect Canadians’ ability to hold their representatives accountable. Because electoral reform would affect voters’ ability to hold politicians accountable for a change in the electoral system, electoral reform should be put to a referendum.

The danger with referendums is that they force citizens to make complicated policy decisions with limited information, understanding of the ramifications of decisions, or consideration of alternative policies. California became the poster case for the over-use of referendums as referendums significantly hurt the quality of governance in that state through the 1990s and 2000s.  Politicians are elected as representatives in order to carefully consider different policy alternatives and to make decisions on our behalf precisely because we as citizens rarely have the time and expertise to make careful decisions on policy details. Politicians are then supposed to be held accountable for these decisions in elections. When they make poor decisions, citizens have the opportunity to vote them out of office. This system does well because it allows the delegation of difficult decisions on issues such as transit and infrastructure spending to individuals who have the time to carefully consider policy options, but also gives them the incentive to act in the public interest by tying their ability to hold on to their jobs to their ability to satisfy the public.

Decisions over electoral reform are different, however, from most policy decisions because they affect the ability of the public to hold politicians accountable. Electoral systems are not neutral in the way that they treat political parties. Different electoral systems benefit different political systems and give parties different incentives to appeal to different groups of citizens. First past the post systems tend to make it easier for centre-right parties to win elections (for more on this see work by Torben Iversen and David Soskice*), alternative vote systems tend to benefit more centrist parties, and proportional systems tend to increase the likelihood that left parties win elections. Politicians and parties thus have strong interests that are independent from the interests of voters. Because electoral systems have a significant impact on the ability of different politicians to remain in office, their biggest concern when debating electoral systems is likely to be the repercussions of each system for their electoral fortunes, as opposed to their benefits and drawbacks with respect to voter representation. The way that politicians weigh the costs and benefits of different electoral systems will in many cases be coloured by the way those electoral system affect their chances of re-election.

On top of this, electoral systems affect the way that politicians are held accountable for their decisions. In first past the post the lower percentage of the vote needed to remain in office means that governments are held accountable by a lower number of voters than they might be in a proportional system. Such systems often give parties governments with around 40% of the vote and thus, if a government can avoid decisions that go against the interests of the 40% of voters that support them, they stand a good chance of remaining in power. In proportional systems governments or governing coalitions need to win 50% of the vote to remain in power and thus need to act in a way that is accountable to the majority of the population. At the same first past the post systems requires that parties in government have support across a large number of ridings and, as a result, forces parties to ensure that they respond to the interests of individuals from variety of different regions. Proportional systems do not force parties to build the same broad based regional coalitions as first past the post systems do, but they do increase electoral importance of smaller but geographically spread out voters such as environmentalists. Changing the electoral system changes the ability of different voters to vote a government out of office, and thus, changes the voters that are most able to hold politicians accountable for their decisions. It is difficult for voters to hold politicians accountable for the electoral the politicians put into place because such an electoral system is likely to weaken the power of voters’ who are hurt by that system.

A popular referendum on electoral reform is valuable because it forces that debate over that reform to be about the way different electoral systems affect the power of different voters as opposed to the way that they affect politicians’ ability to win re-election. In a referendum advocates and opponents of different systems have to justify those systems in a way that appeals to different voters and responds to the how they are affected by different electoral systems. A debate over electoral reform that is carried out as part of a referendum is likely to focus on the implications of different systems for voters who feel that their vote is wasted, voters who are concerned with regional representation, and voters who are worried about the stability of minority and coalition governments. All of these are important trade-offs that should be debated in any discussion of electoral reform. When politicians end up having the final say over a change in the electoral system, it is inevitable that the debate over these trade-off will have to compete with concerns over electoral systems’ affect on different parties’ electoral fortunes. That makes for a weaker electoral reform debate.

A referendum will be a challenge for voters. It will require that individuals learn about the details of different electoral systems and about how they affect the electoral power of different voters. Unfortunately debates over electoral systems require a fair amount of technical knowledge on the part of citizens, while at the same time voters should not trust politicians to act as agents on their behalf when evaluating arguments for and against different electoral systems. The result is that a decision over electoral reform that is legitimate will require a fair amount of citizen education about electoral systems. Only when the decision over an electoral system is taken away from politicians can a decision over electoral reform that takes proper account of voters’ interest be made.

* Iversen, Torben and David Soskice. (2006). “Electoral Institutions and the Politics of Coalitions: Why Some Democracies Redistribute More Than Others.” American Political Science Review. 100(2): 165-181.


Not Much of an Alternative: Not a Whole Lot Would Change Under an Alternative Vote System

The Liberals’ win of the 2015 election means that Canada will see a debate over electoral reform take place over the next 18 months. During the election the Liberals promised that the 2016 election would be the last to take place under Canada’s current first past the post system, and they have now committed to carrying out electoral reform without a referendum. One of the systems under considerations is an alternative vote (also known as a preferential ballot) system. Under a such a system the ridings that exist under first past the post would remain, but candidates would have to win at least 50% of the vote in order to be elected (under the current system candidates only need to win more votes than their closest competitor). Rather than voting for a single candidate voters would rank order candidates. When votes are counted, the last place candidate in a riding is eliminated and anyone who voted for her has their vote redistributed to their second choice candidate. If no candidate has 50% at that point the second last place candidate is eliminated and her votes are redistributed. This continues until a candidate has over 50%.

There are several reasons why individuals my favour electoral reform. Many are frustrated by the ability of parties to win majority governments with less than a majority of the popular vote. The last three majority governments were elected with 39.5%, 39.6%, and 40.9% of the vote. There is also concern that large numbers of votes have no impact the outcome of elections. Votes for 3rd and 4th place candidates, and votes for candidates that win well over 50% of the vote do little to change which candidate or party wins an election. There are also concerns that even the candidates elected to represent particular ridings are not fair representatives because many candidates win in ridings where more than half of the population voted for a different candidates.  Many of these concerns speak to the way that first past the post skews the results of elections by either creating artificial majorities or underepresenting weaker parties.  In this post I look at whether an alternative vote system would have been likely to change the results of the 2011 and 2015 elections. I find that it is unlikely that the results of either election would have been all that different under an alternative vote system. A change to alternative vote would ensure that Members of Parliament would have the support of at least 50% of the population, but do little to change the extent to which parties with less than majority support across the country win majority governments.

In looking at how the 2011 and 2015 elections might have unfolded under alternative vote I use a rather crude method that is based on the results of each election. I first separate out two types of ridings, one in which the winning candidate won at least 50% of the vote and one group in which the winning candidate won at least 45% of the vote. These ridings are unlikely to change hands under an alternative vote system. In the case of the ridings where the top candidate win at least 50%, there would be no need for eliminating last place candidates and redistributing votes. For the candidates with at least 45% of the vote, it is likely they will win enough second place votes from eliminated candidates to put them over the 50% threshold. They need to pick up just 5% (often times less) of the total vote from candidates finishing 3rd or lower. I count both of these groups of ridings as wins for the party that won them under the first past the post system. The remaining ridings, competitive ridings, are ridings that are reasonably likely to change parties under an alternative vote system. For these ridings I look at which parties are competitive in each riding to gain an idea of how much each party could add to the totals they have from the first two groups of ridings. A party is considered competitive in one of these ridings if it finishes first or second, or if it is the third place party and the gap between the 2nd and 3rd place parties is less than the number of votes won by 4th and 5th place parties. All numbers discussed in this post are percentages of total ridings. I use percentages because the number of ridings in the 2015 election was higher than the number of ridings on the 2011 election. All data used in this post comes from Pundits’ Guide.

It should be noted that a drawback of using past results is that campaign tactics and voters’ strategies are likely to be different under different electoral systems. Voters who strategically voted Liberal in 2015, for example, might have ranked NDP candidates first if they were voting in an alternative vote system. For this reason, the numbers in this post should not be taken as exact estimates of what would happened had the 2011 and 2015 elections been run under an alternative vote system. They do provide some idea, though, of what would have happened. It is also unlikely that changes in the voting system would have made a huge difference to electoral outcomes. There is still some incentive to strategic vote in alternative vote systems because the party with the most seats ends up being the one to form government. NDP supporters, for example, might end up voting Liberal to ensure that the Liberals have more seats than the Conservatives, and as a result, keep the Conservatives out of government.

A look at the seats that a party would have locked up even before the competitive seats are taken into account shows the Conservatives in 2011 and the Liberals in 2015 with a large lead in seats. In 2011 the Conservatives had 50% or more of the vote in 35.7% of ridings and 45% or more in an additional 9.4% of ridings. This means that the Conservatives would need to win only an additional 5% from the remaining competitive ridings in order to retain their majority government. In 2015 the Liberals do not do quite as well, but they still win a large number of seats before the competitive seats are taken into account. In 2015 they won 27.2% of seats with at least 50% of the vote, and an additional 13.9% with at least 45% of the vote. This gives the Liberals 41.1% of seats. They need just 9% from the remaining competitive ridings in order to hold on to their majority.

Seats Parties are Likely to Win

Also noteworthy is the small number of safe seats that each of the third parties would have won in an alternative vote system. In 2011 the Liberals had 45% or more in just 2.3% of seats while in 2015 the NDP had 45% or more in just 2.1% of seats. By comparison the NDP in 2011 had 45% or more in 21.1% of seats while the Conservatives in 2015 had 45% or more in 20.1% of seats. The Greens have just one safe seat (Elizabeth May’s) in both years and the Bloc Quebecois have no safe seats in either 2011 or 2015. For one of the third or fourth parties to catch the second place party and become the official opposition they would have to win a large number of the competitive seats. Unless there is a large swing in competitive seats, it is unlikely that either governing party or official opposition in 2011 or 2015 would have changed under an alternative vote system.

A look at the ridings in which parties are competitive shows not much difference between the three largest parties in 2011 and that the Liberals and NDP are competitive in most ridings in 2015. In 2011 each major federal party was competitive in around 20% of ridings. The Conservatives would have needed to win only 23.4% of the ridings in which they were competitive in order to win a majority government. Further, neither the NDP nor the Liberals are competitive in enough ridings to be able to catch the Conservatives and form government. The NDP can win a maximum of 40.9% of seats in 2011 while the Liberals can win a maximum of 24.4% of seats. Neither can exceed the 45.1% of safe seats that the Conservatives had in 2011. Finally, the Liberals would need to win nearly all of their competitive seats and have the NDP win few to none of the seats they are competitive in just to become the official opposition.

Percentage of Seats a Party is Competitive In

In 2015 the Liberals need to win just 27.4% of the seats they are competitive in to retain their majority. The best the Conservatives can do based on the number of seats they are competitive in is 39% while the NDP can win a maximum of 30.2% of seats. Neither could catch the Liberals even if they won none of the seats in which they were competitive in. The NDP could replace the Conservatives as official opposition, but they would need to win an overwhelming number of the seats that they are competitive in while the Conservatives would have to win few of the seats they are in competitive in. Given the discrepancy in safe seats, it is unlikely that the NDP would end up as official opposition if the elections were run using an alternative vote system. In both 2011 and 2015 the opposition and 3rd are not competitive in enough ridings in order to have a strong likelihood of changing the government. In contrast, the relatively small proportion of seats that each governing party would need to win to gain a majority makes such an outcome reasonably likely.

The kinds of ridings that are most likely to be competitive are likely to favour the Liberals. At a centrist party, the Liberals do best in races that are either Conservative-Liberal races or Liberal-NDP races. In the Conservative-Liberal races the Liberals are likely to pick up the 2nd choice votes of New Democrats, while in Liberal-NDP races the Liberals are likely to pick up the second choice votes of Conservatives. In both 2011 and 2015 this suggests that the Liberals will pick up a substantial number of competitive seats. In 2011 15% of seats are either Conservative-Liberal or Liberal-NDP races while in 2015 13% of races are either Conservative-Liberal or Liberal-NDP races. This combined with the likelihood that they will pick up at least some of the other seats suggests that they would get the additional 9% of seats they need in 2015 to hold on to their majority government. The Conservatives might have a slightly more difficult time holding on to their majority in 2011, but the small number of seats that they need makes them likely to do so. There were likely enough Conservative-NDP races and NDP and Green voters that would rank the Conservatives 2nd for the Conservatives to win the additional 5% of seats they need to hold on to their majority in 2011.

Parties Likely to Make it to the Last Round of Counting

Unless alternative vote led to a dramatic change in the way that voters act, the adoption of such a system would have done little to change the results of the 2011 and 2015 elections. It would not have prevented parties that had less than 50% of the popular vote from winning majority governments, nor would it have increased the representation of weaker parties that have difficulty winning seats under first past the post. There are too many ridings in which one party won at least 45% of the vote for alternative vote to have made a substantial difference to the outcome of the 2011 or 2015 elections. A change to a proportional system such as single transferable vote or to mixed member proportional would be needed in order to create such a change in outcomes.